EF Hand Domain Family Member D2 Is Required for T Cell Cytotoxicity Michael Peled, Matthew A

EF Hand Domain Family Member D2 Is Required for T Cell Cytotoxicity Michael Peled, Matthew A. Dragovich, Kieran Adam, Marianne Strazza, Anna S. Tocheva, Irving E. Vega and This information is current as Adam Mor of September 26, 2021. J Immunol published online 1 October 2018 http://www.jimmunol.org/content/early/2018/09/28/jimmun ol.1800839 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2018/09/28/jimmunol.180083 Material 9.DCSupplemental http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 26, 2021 *average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2018 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published October 1, 2018, doi:10.4049/jimmunol.1800839 The Journal of Immunology

EF Hand Domain Family Member D2 Is Required for T Cell Cytotoxicity

Michael Peled,*,† Matthew A. Dragovich,* Kieran Adam,* Marianne Strazza,* Anna S. Tocheva,* Irving E. Vega,‡ and Adam Mor*,x

Programmed cell death 1 (PD-1) is a major coinhibitory receptor and a member of the immunological synapse (IS). To uncover proteins that regulate PD-1 recruitment to the IS, we searched for cytoskeleton-related proteins that also interact with PD-1 using afﬁnity puriﬁcation mass spectrometry. Among these proteins, EF hand domain family member D2 (EFHD2), a calcium binding adaptor protein, was functionally and mechanistically analyzed for its contribution to PD-1 signaling. EFHD2 was required for PD-1 to inhibit cytokine secretion, proliferation, and adhesion of human T cells. Interestingly, EFHD2 was also required for human T cell– mediated cytotoxicity and for mounting an antitumor immune response in a syngeneic murine tumor model. Mechanistically,

EFHD2 contributed to IS stability, lytic vesicles trafﬁcking, and granzyme B secretion. Altogether, EFHD2 is an important Downloaded from regulator of T cell cytotoxicity and further studies should evaluate its role in T cell–mediated inﬂammation. The Journal of Immunology, 2018, 201: 000–000.

he immunological synapse (IS) is the interface between Most of the immune checkpoints are coinhibitory receptors that T cells and APC. The IS consists of segregated clusters of directly inhibit signaling downstream of the TCR. T proteins involved in T cell activation. A key member of Programmed cell death 1 (PD-1) is a typical coinhibitory re- http://www.jimmunol.org/ these clusters is the TCR that is engaged by an Ag-loaded MHC ceptor,expressedonactivatedTcellsandclustersintheIS on the APC. Productive interactions initiate signaling cascades alongside the TCR upon engagement by its ligands, PD-1 ligand 1 that result in T cell adhesion, proliferation, and cytokine secretion. (PDL1) or PDL2 (4). PDL1 is widely expressed on both hema- When the IS is composed of a cytotoxic T cell and a corresponding topoietic and nonhematopoietic cells, including tumor cells, target cell (such as a tumor or virally infected cell), lytic granules whereas PDL2 is mainly expressed on APC (5–9). Upon ligation, are released to eliminate the counter cells. To execute these func- PD-1 recruits the phosphatase SHP2 to dephosphorylate phos- tions, a stable synapse between the T cell and the target cell must phoproteins downstream of the TCR and to inhibit T cell acti- be formed. Indeed, TCR-induced synapses are characterized by vation (10–15). PD-1 knockout mice develop inflammatory reorientation of the microtubule-organizing center and substantial cardiomyopathy and glomerulonephritis (16, 17), and PD-1 block- by guest on September 26, 2021 actin accumulation at the boundaries of the contact region (1–3). ade induces potent antitumor immune response in patients with Whereas T cell functions are required for mounting protective various types of malignancies (18, 19). immune responses, it can also result in autoimmunity unless tightly We have recently used affinity purification mass spectrometry regulated. Accordingly, immune responses are restrained by a (MS) to discover proteins, besides SHP2, that interact with the plethora of mechanisms, including thymic clonal deletion of self- intracellular tail of PD-1 and regulate its function (14). Because reactive T cells, inhibitory cells such as regulatory T cells, and anti- recruitment of PD-1 to the IS is critical for its function (4), we inflammatory cytokines (i.e., IL-10 and TGF-b). Another important used this data set to screen for proteins that interact with PD-1 and mechanism to restrict immune responses is immune checkpoints. that also regulate IS formation through cytoskeleton organization. Multiple proteins were uncovered, including Swiprosin-1/EF hand domain family member D2 (EFHD2). EFHD2 knockdown (KD) *Perlmutter Cancer Center, New York University School of Medicine, New York, NY 10016; †Pulmonary Department, The Chaim Sheba Medical Center, Ramat Gan blocked PD-1 inhibitory functions in T cells through inhibition of 52620, Israel; ‡Department of Translational Science and Molecular Medicine, Col- stable IS formation. However, EFHD2 was also required for lege of Human Medicine, Michigan State University, Grand Rapids, MI 49503; and x T cell–mediated cytotoxicity and for antitumor immune response, Columbia Center for Translational Immunology, Columbia University Medical Center, New York, NY 10032 implying a pervasive role for this protein in IS formation. ORCID: 0000-0003-3106-1183 (M.S.). Received for publication June 18, 2018. Accepted for publication August 24, 2018. Materials and Methods This work was supported by a grant from the National Institutes of Health (R01 General reagents AI25640). RPMI 1640 medium, DMEM, Dulbecco’s PBS, and FBS were purchased Address correspondence and reprint requests to Dr. Adam Mor, 630 W 168 Street, from Life Technologies. Opti-MEM I was purchased from Invitrogen. PH8-406, Columbia University Medical Center, New York, NY 10032. E-mail Ficoll–Paque was purchased from GE Healthcare. Staphylococcus en- address: [email protected] terotoxin E (SEE) was acquired from Toxin Technology. BCA assay was The online version of this article contains supplemental material. purchased from Pierce Biotechnology. Poly-L-lysine, puromycin, and fibro- Abbreviations used in this article: EFHD2, EF hand domain family member D2; IS, nectin were obtained from Sigma-Aldrich. immunological synapse; ITSM, immunoreceptor tyrosine-based switch motif; KD, knockdown; MS, mass spectrometry; MTS, Aqueous One Solution Cell Proliferation; Cell culture, transfection, and stimulation NYU, New York University; PD-1, programmed cell death 1; PDL1, PD-1 ligand 1; SEE, Staphylococcus enterotoxin E; siRNA, small interfering RNA; TIL, tumor- Primary human T cells were isolated from whole blood using RosetteSep infiltrating lymphocyte; 39-UTR, 39-untranslated region; WT, wild type. (STEMCELL Technologies). Cells were maintained in enriched media at 5% CO2 at 37˚C. Primary murine T cells were isolated from spleens of Copyright Ó 2018 by The American Association of Immunologists, Inc. 0022-1767/18/$37.50 10- to 12-wk-old mice, followed by negative selection using Dynabeads

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1800839 2 EFHD2 IS REQUIRED FOR T CELL CYTOTOXICITY

(Invitrogen) T cell isolation kit. Jurkat T cells and Raji B cells were Flow cytometry obtained from the American Type Culture Collection and maintained in RPMI 1640 medium supplemented with 10% FBS and 100 U/ml peni- T cells were stained with fluorescently conjugated Abs specific for PD-1 in cillin and streptomycin. MC38 cells were provided by Benjamin Neel FACS Buffer, washed, and fixed in 1% paraformaldehyde. Events were (New York University [NYU]) and maintained in RPMI 1640 medium recorded using FACSCalibur (BD Biosciences) and analyzed using FlowJo supplemented with 10% FBS and 100 U/ml penicillin and streptomycin. software. To obtain single-cell suspensions from tumors, tissues were HEK 293T cells were obtained from the American Type Culture Col- digested and cells were blocked with anti-FcR prior to staining with Abs against CD3, PD-1, CD44, CD8, CD4, and CD45 (BioLegend). For anti- lection and maintained in 5% CO2 at 37˚C in DMEM media supplemented with 10% FBS and 100 U/ml penicillin and streptomycin. DNA Foxp3 (BioLegend) staining, cells were fixed and permeabilized using expression constructs were introduced into the cells by nucleofection True-Nuclear Transcription Factor Buffer Set (BioLegend). Samples were (Lonza Bioscience) with efficiency of 50–70%. Cells were stimulated collected on a BD LSR II Flow Cytometer (BD Biosciences), and data with magnetic beads (ratio of 1:3 cells per bead), which were conjugated were analyzed using FlowJo software. with the following protein combinations (the ratio in parentheses Static adhesion assay indicates the relative concentration of each protein): anti-CD3/IgG1 (1:3), anti-CD3/PDL1-Fc/IgG1 (1:2:1), anti-CD3/PDL2-Fc/IgG1 (1:2:1), Static T cell adhesion to fibronectin-coated plates was performed as anti-CD3/anti-CD28/IgG1 (1:1:2), or anti-CD3/anti-CD28/PDL1-Fc reported earlier (21). (1:1:2). Constructs were introduced into HEK 293T cells using Super- Fect Transfection Reagent (Qiagen) according to the manufacturer’s Conjugate formation assay protocol. SEE (Toxin Technology)-loaded Raji B cells (1.5 3 106) were mixed with 6 Abs Jurkat T cells (1.5 3 10 ), plated on poly-L-lysine (5 mg/ml)–coated glass bottom culture 35-mm plates (ibidi), and subjected to brief centrifugation

Anti-CD3 (UCHT1), PDL1-Fc, and PDL2-Fc were purchased from R&D prior to imaging. Images were taken with Zeiss 700 confocal microscope Downloaded from Systems. IgG1 (isotype control) was purchased from Jackson Immuno- and quantiﬁed for conjugate formation as previously described (10, 22). Research. Anti-CD28 (CD28.2) was purchased from eBioscience. Anti-GFP mAb-agarose (MBL International), anti-GFP (Invitrogen), rat anti-mouse Mice CD16/CD32 (Biosciences), and anti-EFHD2 (Abcam) Abs were used as described. Alexa 488 anti-mouse CD3, percp/Cy5.5 anti-mouse CD8a, Animal studies were approved by the NYU Institutional Animal Care and Allophycocyanin/Cy7 anti-mouse CD45, PE anti-mouse CD4, Alexa 647 Use Committee. EFHD2 knockout mice were generated by Dr. I. Vega anti-mouse Foxp3, PE/Cy7 anti-mouse CD279 (PD-1), BV421 anti- (Michigan State University). Brieﬂy, disruption of EFHD2 gene in em-

bryonic stem cells was performed by homologous recombination using a http://www.jimmunol.org/ mouse/human CD44, violet 421 IgG2b isotype control, PE/Cy7 IgG2b replacement-type targeting vector (Neo-LacZ). Embryonic stem cells were isotype control, and Alexa 647 IgG2b isotype control Abs were purchased from BioLegend and used as described. electroporated with neomycin (G418)-LacZ clones. Chimeric males were obtained and bred with C57BL/6 wild type (WT) females for 10 genera- DNA constructs tions. For the syngeneic tumor model, mice were injected on day 0 with 1 3 106 MC38 cells s.c. in the right flank. Tumor diameter was measured pMSCV-PD-1-YFP was a generous gift from J. Allison (MD Anderson every 3 d with an electronic caliper and reported as volume using the Cancer Center). PD-1–GFP fusion expression constructs were generated formula (width2 3 length) 3 p / 6 (23). Mice were sacrificed when tumors through PCR amplification and cloning into pGFP-N1 vector. Residues reached the size of 2000 mm3. Tumor-infiltrating lymphocytes (TIL) were 192–288 of the PD-1 were used for cloning the tail of PD-1. pEFHD2- analyzed as described earlier. Cherry was generated by cloning of the EFHD2 gene from pDONR221- EFHD2 (DNASU) into mCherry-hLC3B-pcDNA3.1 [D. Rubinsztein (20); Cytotoxicity assay by guest on September 26, 2021 AddGene no. 40827]. mCherry-Lifeact-7 was a gift from M. Davidson (AddGene no. 54491). pMD2G and psPAX2 were gifts from M. Philips Murine splenocytes were cultured with 1 mg/ml SEE and 1000 U/ml (NYU). murine IL-2 (Miltenyi Biotec) for 72 h, followed by isolation of CD8 T cells using isolation kit (Miltenyi Biotec). SEE-loaded Raji B cells Knocking down of EFHD2 (target cells) were mixed with CD8 T cells at different ratios, as indicated, and cytotoxicity was tested using lactate dehydrogenase cytotoxicity EFHD2 was stably knocked down in Jurkat T cells by RNA interference kit (Thermo Fisher Scientific). In some experiments, CD8 T cells were using Mission shRNA plasmids (Sigma-Aldrich). Lentiviral particles were stained with Lysotracker (Thermo Fisher Scientific) before mixing with generated by transfecting HEK 293T cells with pMD2G, psPAX2, and the Raji cells, being plated on poly-L-lysine (5 mg/ml)–coated glass bottom shRNA plasmid using SuperFect (Qiagen). T cells were transduced by cen- culture 35-mm plates (ibidi), and being subjected to brief centrifugation trifugation and selected with puromycin. SMARTpool ON-TARGETplus prior to imaging. Images were taken with Zeiss 700 confocal microscope EFHD2 and nontargeting control small interfering RNA (siRNA) (Dharma- and analyzed with ImageJ (National Institutes of Health). con) were used according to the manufacturer’s instruction. For rescue experiments, a 39-untranslated region (39-UTR) targeting siRNA was used. Histopathology m ELISA Tumors were fixed in 10% formalin, embedded in paraffin, cut in 5- m thick sections, and stained with H&E or anti-CD3 (Abcam). Bound Abs To determine the concentration of secreted proteins after stimulation, hu- were detected with peroxidase-based staining. man and mouse IL-2 (BioLegend), human IFN-g (BioLegend), and mouse Granzyme B (Invitrogen) kits were used according to the manufacturer’s Statistics protocols. Cells were stimulated as indicated. Values are reported as means 6 SEM. Statistical analyses were performed using Student t test and ANOVA analysis. All statistical analyses were Cell proliferation assay performed using GraphPad Prism (version 6.0). Cell proliferation assay was performed using tetrazolium compound based CellTiter 96 Aqueous One Solution Cell Proliferation (MTS) assay Results (Promega). T cells were activated with Ab-coated beads and cultured for Multiple IS-related proteins interact with the tail of PD-1 4–6 d. MTS cell viability assay was performed as indicated, according to the manufacturer’s instructions. To identify proteins that interact with the tail of PD-1 and that are also related to IS formation and stability, we used our previously Immunoprecipitation, MS, and Western blot analysis published dataset of the PD-1 interactome (14). As described, GST- Cell lysates were mixed with anti-GFP mAb coupled to agarose beads to tagged PD-1 cytoplasmic tail was used to affinity purify inter- enrich for GFP tagged proteins according to the manufacturer’s protocols acting proteins that were subsequently identified by MS. A total (MBL International). Pull down lysates were separated by Tris-glycine PAGE, transferred to nitrocellulose filters, and visualized as previously of 813 proteins that interacted with GST itself were excluded described (14). Affinity purification MS with GST-tagged PD-1 tail was (Fig. 1A). Of the 220 proteins that specifically interacted with PD- previously described (14). 1, we considered candidates that were annotated as related to the The Journal of Immunology 3

FIGURE 1. Multiple IS-related proteins interact with the tail of PD-1. (A) Venn diagram showing the numbers of proteins that interact with GST or with GST–PD-1. (B) Functional pathways enrichment analysis of the proteins that interact speciﬁcally with GST–PD-1 and are related to the IS. The top nine signiﬁcantly enriched functional clusters (y-axis) are displayed over the corresponding 2log10 p value (x-axis). The p value represents the probability of discovering at least x number of genes (x = number in brackets) out of the total number of the genes in the list (n = 19, Table I) annotated to a particular functional cluster given the proportion of genes in the whole genome that are annotated to that functional cluster. Downloaded from

IS by performing the following query in the GeneCards database cells by PDL1-expressing Raji cells loaded with SEE, both PD-1 (www.genecards.org) (24): “[phenotypes] (immune) AND [all] and EFHD2 polarized toward the IS (Fig. 3B; lower panel), sup- (cytoskeleton synapse).” Accordingly, 19 PD-1–binding proteins porting functional association. were identified (Table I). Pathways analysis (25) confirmed in- Cytoskeleton organization through actin polymerization is re- volvement of these proteins in IS-related functions (Fig. 1B), quired to establish stable IS and for proper recruitment of key http://www.jimmunol.org/ supporting the biological relevance of the identified proteins. coreceptors (27), including PD-1. A previous study has shown that EFHD2 was localized to areas of actin polymerization (28) and EFHD2 is required for PD-1 inhibitory functions that EFHD2 could promote actin polymerization; hence, we Among the proteins that interact with PD-1 (Table I), EFHD2 is of knocked down EFHD2 in primary human T cells and recorded the particular interest due to the shared autoimmune characteristics of localization of GFP–PD-1. As shown, in the absence of EFHD2, EFHD2 and PD-1–deficient mice (26). Coimmunoprecipitation PD-1 failed to polarize toward the IS (Fig. 3C), indicating that confirmed physical interaction between GFP–PD-1 and endoge- EFHD2 is needed for PD-1 subcellular localization. nous EFHD2 in Jurkat T cells (Fig. 2A). Knocking down EFHD2 To further study the contribution of EFHD2 to synapse for- by guest on September 26, 2021 in primary human T cells with inhibitory RNA targeting the mation, we cultured primary human T cells deficient of EFHD2 39-UTR of this gene (Supplemental Fig. 1A) resulted in near with Raji B cells loaded with SEE and recorded the tightness of the complete elimination of the inhibitory effect of PD-1 on IL-2 synapses (Supplemental Fig. 2B). Shortly after coculturing, 59% secretion downstream the Ag receptor (Fig. 2B). This phenotype of the control cells formed stable conjugates (Fig. 3D), whereas was rescued by expressing EFHD2 lacking the 39-UTR (Fig. 2B). for the EFHD2 KD cells, only 24% of the cells maintained this Importantly, PD-1 expression levels were maintained in the ab- phenotype, suggesting that EFHD2 is required to maintain stable sence of EFHD2 (Supplemental Fig. 1B). EFHD2 was required IS. During synapse maturation, actin is depleted from the center of for PD-1 inhibitory functions whether PDL1 or PDL2 was used to the synapse (Supplemental Fig. 2C) (29) and accumulates at the engage PD-1 (Fig. 2B). EFHD2 was also required for mediating PD-1 inhibitory signals when T cells were stimulated by cross- linking the coreceptor CD28 (Fig. 2C). KD of EFHD2 blocked Table I. Immune-, cytoskeleton-, and synapse-related interactions of the PD-1 tail PD-1–induced ERK dephosphorylation (Supplemental Fig. 1C). Functionally, EFHD2 was necessary to mediate the inhibitory Gene UniProt ID Peptide Spectrum Match effects of PD-1 on IFN-g secretion (Fig. 2D) and adhesion of T cells to fibronectin-coated wells (Fig. 2E). The ability of PD-1 PFN1 P07737 20 to inhibit IL-2 secretion (Fig. 2F) and proliferation (Fig. 2G) PTPN11 Q06124 17 GRB2 P62993 15 downstream of the TCR was also dependent on EFHD2 in primary MYO5A Q9Y4I1 13 mouse T cells isolated from EFHD2-deficient mice. Altogether, TLN1 Q9Y490 11 EFHD2 is an essential mediator of PD-1 signaling and functions in EFHD2 Q96C19 7 both human and murine T cells in vitro. PDIA3 P30101 7 VAV1 P15498 5 EFHD2 colocalizes with PD-1 in the IS ATP2A2 P16615 5 ANXA1 P04083 5 To explore whether EFHD2 and PD-1 are localized to the same RPL22 P3528 5 subcellular compartment in the context of IS formation, we EVL Q9UI08 4 overexpressed Cherry-EFHD2 in Jurkat T cells. EFHD2 protein CAPN1 P07384 4 was distributed to the cytosolic compartment (Fig. 3A; right panel DOCK2 Q92608 4 PSMC4 P43686 3 and Supplemental Fig. 2A), whereas GFP–PD-1 (Supplemental ELMO1 Q92556 2 Fig. 1D) was localized to the plasma membrane and to a vesicu- ITK Q08881 2 lar compartment (Fig. 3A; left panel). When we overexpressed RAF1 P04049 2 both proteins in the same cell, EFHD2 was partly enriched at the DYNLL1 P63167 2 plasma membrane (Fig. 3B; upper panel), but upon activating the ID, identifier. 4 EFHD2 IS REQUIRED FOR T CELL CYTOTOXICITY Downloaded from http://www.jimmunol.org/

FIGURE 2. EFHD2 is required for PD-1 inhibitory functions. (A) Jurkat T cells were transiently transfected with GFP–PD-1 or GFP alone, whole-cell by guest on September 26, 2021 lysates (WCL) were incubated with anti-GFP Ab, and immunoprecipitated proteins were detected by immunoblotting, as indicated. A representative blot of three independent experiments is shown. (B–D) Freshly isolated human CD3 T cells were transfected with nontargeting control siRNA (scrambled) or with siRNA targeting EFHD2 or with EFHD2 expression plasmids, as indicated. Twenty-four hours later, cells were stimulated with magnetic beads coated with anti-CD3, anti-CD3 + PDL2, anti-CD3 + PDL1, anti-CD3 + CD28, or anti-CD3 + CD28 + PDL1, as indicated. After additional 24 h, media was harvested and IL-2 or IFN-g levels were measured by ELISA. EFHD2 over expression (OE) was rescued by expressing EFHD2 lacking the 39-UTR. (E) Jurkat T cells, stably transfected with shRNA targeting EFHD2 or a control nontargeting shRNA (scrambled), were stimulated as indicated and subjected to static adhesion assay to fibronectin-coated wells. (F) Freshly isolated mouse T cells from WT mice or EFHD2 knockout (EFHD22/2) mice were stimulated with magnetic beads coated with anti-CD3 or anti-CD3 + PDL2 for 48 h. IL-2 levels in the media were measured by ELISA. (G) Freshly isolated mouse T cells from WT mice or EFHD2 knockout mice were stimulated with magnetic beads, as indicated, and cell proliferation was monitored using the MTS assay. Values are expressed as increase in OD compared with day 0. All the experiments were done at least three times (n $ 3). *p , 0.05, **p , 0.001, ***p , 0.0001, unpaired Student t-test. periphery of the contact area between the cells. Actin visualization between the experimental groups (Fig. 4C, 4D). Flow analysis by fluorescently tagged Lifeact peptide (30) revealed F-actin revealed that both CD4 and CD8 TIL collected from EFHD2-null clearance from the center of the synapse in the control cells but mice were more activated compared with the control group, as not in the EFHD2-deficient cells (Fig. 3E). Altogether, these re- measured by the expression of both PD-1 and the activation sults suggest that EFHD2 is required for the formation and the marker CD44 (Fig. 4E). Cells isolated from spleens of the same maturation of the synapse, which affects PD-1 recruitment to the animals did not demonstrate higher percentage of activated T cells same compartment. (Fig. 4E). There was also no significant difference in the percentage of regulatory T cells among the TIL to explain the in- EFHD2 is necessary for antitumor immune response in vivo creased tumor growth rate in the EFHD2-null mice (Fig. 4F). PD-1–deficient mice demonstrate reduced tumor growth because of increased antitumor T cell activity (23). Because EFHD2 is EFHD2 is required for T cell–induced cytotoxicity necessary for PD-1 cellular functions and polarization, we antic- Although EFHD2 was necessary for PD-1 signaling and cellular ipated that tumor growth rate in EFHD2-null mice would be re- functions (Fig. 2), it was also required for stable IS formation duced. To test that, MC38 murine colon adenocarcinoma cells (Fig. 3), a critical constituent for cytotoxic T cell functions. We were implanted s.c. in EFHD2 knockout mice, but unexpectedly, hypothesized that cytotoxicity in EFHD2-deficient cells is reduced tumor growth rates were increased in the null mice compared with compared with the control cells, explaining the increased tumor the WT littermates (Fig. 4A, 4B). growth rate in the former (Fig. 4). To assess that, we used isolated Quantification and organization of TIL failed to explain this primed CD8 T cells from both WT and EFHD2-null mice unpredicted phenotype in the absence of significant differences and subjected them to super Ag–induced cytotoxicity assay (31). The Journal of Immunology 5 Downloaded from http://www.jimmunol.org/

FIGURE 3. EFHD2 colocalizes with PD-1 in the IS. (A) Jurkat T cells were transfected with the indicated expression plasmids and imaged by confocal microscopy. Images are representative of at least 50 cells from each of three independent experiments. The percentage of the dominant phenotype for each by guest on September 26, 2021 condition is shown. Scale bars, 10 mm. (B) Jurkat T cells were transfected with the indicated expression vectors prior to coculturing with Raji B cells that either express or not the ligand PDL1 and are loaded with SEE. Images are representative of at least 50 cells from each of three independent experiments. The percentage of the dominant phenotype for each condition is shown. Scale bars, 10 mm. Line analysis shows fluorescent intensities of EFHD2 along the diameter of the cell. (C) Freshly isolated human T cells were transfected with nontargeting siRNA (siControl) or siRNA targeting EFHD2 (siEFHD2) and with GFP–PD- 1 expression plasmid, followed by coculturing with Raji B cells expressing PDL1 and loaded with SEE. Cells were subjected to real-time imaging by confocal microscopy. Images are representative of at least 50 cells from each of three independent experiments. The percentage of the dominant phenotype for each condition is shown. Scale bars, 10 mm. Line analysis shows fluorescent intensities of GFP–PD-1 along the diameter of the cell. (D) Percentage of transient and stable conjugates. Mean of five independent experiments is shown. (E) Freshly isolated human T cells were transfected with Cherry LifeAct and with nontargeting siRNA (siControl) or siRNA targeting EFHD2 (siEFHD2), followed by coculturing with Raji B cells expressing PDL1 and loaded with SEE. Images are representative of at least 50 cells from each of four independent experiments. Arrows indicate actin accumulation. The percentage of the dominant phenotype for each condition is shown. Scale bars, 10 mm.

EFHD2 was required for effective killing of target cells in vitro Because EFHD2-deficient mice show unregulated responses to (Fig. 5A). Even though EFHD2-null lymphocytes were more ac- infectious stimuli and spontaneously develop autoantibodies tivated (Fig. 4E), their overall ability to counter tumor cells was (26), and based on our finding that EFHD2 is necessary for PD-1 limited (Fig. 5A). Mechanistically, polarization of lytic granules function in vitro, we speculated that EFHD2-null mice would toward the synapse in the EFHD2-null T cells was impaired show resistance to immunogenic tumors (23). However, MC38 compared with WT cells (Fig. 5B, 5C). In addition, lower levels of tumors exhibited an increased growth rate in the EFHD2- the enzyme Granzyme B were released from EFHD2-deficient deficient mice despite an increased expression of cellular activa- cells compared with WT control cells (Fig. 5D). tion markers. This unexpected discrepancy is explained, at least in part, by the ability of EFHD2 to control additional key cellular Discussion functions, such as IS formation and cytotoxicity, besides mediat- PD-1 polarization and trafficking to the IS are necessary for its ing PD-1 signaling and polarization. function as an inhibitory immune checkpoint (4). To uncover PD-1 EFHD2 is a 240 aa calcium sensor protein that was originally interacting partners that are also required for its localization, we described in lymphocytes (32). It consists of an N-terminal region explored our previous dataset (14) through GeneCard’s func- of low complexity with an alanine stretch, a functional SH3- tional annotation (24). We found that EFHD2, a calcium-binding binding motif, two functional EF-hand domains, and a C-terminal adaptor protein (32–35), physically interacted with PD-1 and coiled-coil domain (Supplemental Fig. 3) (36). The mechanism of also colocalized with it at the IS. EFHD2 was necessary not just the interaction between PD-1 and EFHD2 is unclear. EFHD2 con- for PD-1 trafficking to the IS but also for PD-1 function in vitro. tains SH3-binding motifs; however, the tail of PD-1 lacks SH3 6 EFHD2 IS REQUIRED FOR T CELL CYTOTOXICITY Downloaded from http://www.jimmunol.org/ by guest on September 26, 2021

FIGURE 4. EFHD2 is necessary for antitumor immune response in vivo. MC38 cells were inoculated into EFHD2-deficient male mice (A) or female mice (B) and into WT littermates. After tumor inoculation (day 0), tumor volume was measured as described in the method section. Results are expressed as mean 6 SEM (n = 4 per group). (C) Immune histochemical staining in two representative tumor sections. Original magnification 320. (D–F) Single-cell suspensions from tumor tissues and spleens were analyzed by fluorescence-activated cell sorter. Results are shown as percentage of indicated subset in particular immune cell population. (D) Percentage of CD45/CD3+ cells of total tumor cells. (E) Expression of CD44 and PD-1 on splenocytes and tumor- infiltrating CD8 and CD4 cells from MC38 tumor-bearing mice. Bar graphs representing percentages of the CD44/PD-1–expressing populations are shown at the bottom. (F) Flow cytometric analysis depicting percentages of Foxp3-positive cells on tumor-infiltrating CD4 T cells from MC38 tumor-bearing mice. Bar graphs representing percentages of the CD4/Foxp3–expressing population is shown on the right. All the experiments were done at least three times (n $ 3). *p , 0.05, **p , 0.001. and instead contains phosphotyrosine-based motifs that can bind found to be a PD-1–binding partner (Table. I) (14). GRB2 contains to SH2 domains, ITIM, and the immunoreceptor tyrosine-based both SH2 and SH3 domains and thus could potentially link PD-1 switch motif (ITSM) (14). EFHD2 was affinity purified by both and EFHD2. Whereas we and others have shown that the ITIM and WT PD-1 tail and by the PD-1 Y223F, Y248F tail, which lacks ITSM of PD-1 are necessary for SHP2 binding and activation (12, the ITIM and the ITSM (14). Thus, its interaction with PD-1 ap- 14), there could be an early and transient interaction between PD-1 pears to be tyrosine independent. Remarkably, GRB2 was also and GRB2, which is later taken over by SHP2. Importantly, it was The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/

FIGURE 5. EFHD2 is required for T cell–induced cytotoxicity. (A) WT control and EFHD22/2 CD8 T cells were used as effector cells toward SEE- loaded Raji target cells in 4 h cytotoxicity lactate dehydrogenase release assay. (B) A representative image of the localization of lytic granules in CD8 T cells from WT and EFHD22/2 mice following coculturing with Raji target cells. Arrows indicated granule localization. Scale bars, 10 mm. (C) Distance of granules from the contact area between the cells within 5 min of synapse formation. Fifty to seventy individual granules were counted from multiple cells acquired through three independent experiments. Data represent the mean 6 SEM. (D) CD8 T cells from either WT or EFHD22/2 were cocultured with SEE-loaded Raji target cells, and levels of released Granzyme B were measured by ELISA. (n = 3). *p , 0.05. by guest on September 26, 2021 previously demonstrated that the ITIM and ITSM of PD-1 are not participates in a negative feedback loop, preventing NF-kB required for its clustering (4), implying that an alternative mecha- signaling as a result of BCR activation (36). Others have shown nism should be considered. opposing results, in which EFHD2 augmented upregulation of A previous report has already demonstrated that EFHD2 was NF-kB target genes in PMA/ionomycin–stimulated HMC-1 localized to areas of actin dynamics (28). EFHD2 was localized to cells (40). Additional discrepancy is demonstrated by experi- the microvilli-like protrusions in HMC-1, a mast cell leukemia mental data using the EFHD2-null mice. These mice elicit en- cell line (37). Moreover, and in agreement with our findings, hanced Th2 cell–mediated IgE and IgM production as well as EFHD2 was recruited to the actin-rich region in Jurkat T cells enhanced germinal center formation in response to infection during T cell activation, and overexpression of EFHD2 enhanced with the helminth Nippostrongylus brasiliensis as compared lamellipodia formation and cell spreading, whereas KD of this with WT mice, suggesting that EFHD2 acts as a negative reg- protein resulted in an opposite effect (28). Mechanistically, it was ulator of type 2 immunity (26). However, LPS-induced sepsis shown that EFHD2 induced actin bundling in the presence of resulted in higher mortality, severe organ dysfunction, restrained calcium. Our findings also support a role for EFHD2 in actin macrophage recruitment in affected organs, and attenuated inflam- dynamics, demonstrated by the reduction in the number of stable matory cytokine production in EFHD2-null mice compared with the synapses in the EFHD2-deficient cells and by reduced actin WT littermates, implying a proinflammatory role of EFHD2. This recruitment to the periphery of the synapses. The lack of PD-1 was attributed to impaired bactericidal capacity and decreased trafficking to the IS and the reduced polarization of cytotoxic HLA-DR expression in EFHD2-null macrophages, secondary granules are likely direct consequences of altered actin dynamics to EFHD2 regulation of IFN-g receptors (41). Thus, EFHD2 secondary to EFHD2 depletion. pleotropic effect in the innate and adaptive immune response The role of EFHD2 in innate and adaptive immune responses requires further study. was previously studied. However, it was not clear what was the net To summarize, our findings show that EFHD2 is necessary for effect of EFHD2 in different immune cells, in terms of activation or both anti-inflammatory and proinflammatory functions of T cells. suppression of their function. Kwon et al. (28) reported that This is due to its generalized role in regulating both PD-1 signaling EFHD2 enhanced SDF-1–mediated T cell spreading, and Kroczek and trafficking but also in mediating adhesion, cytotoxic granules et al. (38) showed that murine EFHD2 was a positive regulator of polarization, and T cell–mediated cytotoxicity. The net effect Syk activity in response to BCR stimulation in the murine cell line in vivo, at least in the MC38 tumor model, points toward a WEHI231. Whereas these studies imply a proinflammatory role proinflammatory role for EFHD2, as tumor growth rate in for EFHD2, it was also found to block the expression of the EFHD2-deficient mice was enhanced, supporting data from the antiapoptotic NF-kBtargetgeneBCL-XL (39). It has been LPS-induced sepsis experiments (41). It is important to note that proposed that EFHD2 binds calcium directly (33) and therefore both the MC38 model and the LPS-induced sepsis model are Th1 8 EFHD2 IS REQUIRED FOR T CELL CYTOTOXICITY directed models (unlike the helminth infection model, which is a 18. Couzin-Frankel, J. 2013. Breakthrough of the year 2013. Cancer immunother- apy. Science 342: 1432–1433. Th2 immune response). Thus, EFHD2 may have different functions 19. Tumeh, P. C., C. L. Harview, J. H. Yearley, I. P. Shintaku, E. J. Taylor, L. Robert, depending on the stimulus and the course of the immune response, B. Chmielowski, M. Spasic, G. Henry, V. Ciobanu, et al. 2014. PD-1 blockade and targeting EFHD2 in Th1-driven autoimmune disorders may induces responses by inhibiting adaptive immune resistance. Nature 515: 568– 571. have therapeutic benefit. 20. Jahreiss, L., F. M. Menzies, and D. C. Rubinsztein. 2008. The itinerary of autophagosomes: from peripheral formation to kiss-and-run fusion with lyso- Acknowledgments somes. Traffic 9: 574–587. We thank the NYU Proteomics Laboratory for helping with the mass spec- 21. Strazza, M., I. Azoulay-Alfaguter, A. Pedoeem, and A. Mor. 2014. Static adhesion assay for the study of integrin activation in T lymphocytes. J. Vis. Exp. trometry experiment. DOI: 10.3791/51646. 22. Strazza, M., I. Azoulay-Alfaguter, B. Dun, J. Baquero-Buitrago, and A. Mor. 2015. CD28 inhibits T cell adhesion by recruiting CAPRI to the plasma mem- Disclosures brane. J. Immunol. 194: 2871–2877. The authors have no financial conflicts of interest. 23. Woo, S. R., M. E. Turnis, M. V. Goldberg, J. Bankoti, M. Selby, C. J. Nirschl, M. L. Bettini, D. M. Gravano, P. Vogel, C. L. Liu, et al. 2012. Immune inhibitory molecules LAG-3 and PD-1 synergistically regulate T-cell function to promote References tumoral immune escape. Cancer Res. 72: 917–927. 24. Safran, M., V. Chalifa-Caspi, O. Shmueli, T. Olender, M. Lapidot, N. Rosen, 1. Stinchcombe, J. C., M. Salio, V. Cerundolo, D. Pende, M. Arico, and M. Shmoish, Y. Peter, G. Glusman, E. Feldmesser, et al. 2003. Human gene- G. M. Griffiths. 2011. Centriole polarisation to the immunological synapse di- centric databases at the Weizmann institute of science: GeneCards, UDB, CroW rects secretion from cytolytic cells of both the innate and adaptive immune 21 and HORDE. Nucleic Acids Res. 31: 142–146. systems. BMC Biol. 9: 45. 25. Chen, J., E. E. Bardes, B. J. Aronow, and A. G. Jegga. 2009. ToppGene suite for 2. Ueda, H., M. K. Morphew, J. R. McIntosh, and M. M. Davis. 2011. CD4+ T-cell gene list enrichment analysis and candidate gene prioritization. Nucleic Acids synapses involve multiple distinct stages. Proc. Natl. Acad. Sci. USA 108: Downloaded from Res. 37: W305-11. 17099–17104. 26. Brachs, S., A. Turqueti-Neves, M. Stein, D. Reimer, B. Brachvogel, M. Bo¨sl, 3. Das, R., H. Bassiri, P. Guan, S. Wiener, P. P. Banerjee, M. C. Zhong, A. Veillette, T. Winkler, D. Voehringer, H. M. Ja¨ck, and D. Mielenz. 2014. Swiprosin-1/ J. S. Orange, and K. E. Nichols. 2013. The adaptor molecule SAP plays essential EFhd2 limits germinal center responses and humoral type 2 immunity. Eur. J. roles during invariant NKT cell cytotoxicity and lytic synapse formation. Blood Immunol. 44: 3206–3219. 121: 3386–3395. 27. Kumari, S., S. Curado, V. Mayya, and M. L. Dustin. 2014. T cell antigen receptor 4. Yokosuka, T., M. Takamatsu, W. Kobayashi-Imanishi, A. Hashimoto-Tane, Biochim. Biophys. Acta M. Azuma, and T. Saito. 2012. Programmed cell death 1 forms negative co- activation and actin cytoskeleton remodeling. 1838: stimulatory microclusters that directly inhibit T cell receptor signaling by 546–556. recruiting phosphatase SHP2. J. Exp. Med. 209: 1201–1217. 28. Kwon, M. S., K. R. Park, Y. D. Kim, B. R. Na, H. R. Kim, H. J. Choi, I. Piragyte, http://www.jimmunol.org/ 5. Dong, H., G. Zhu, K. Tamada, and L. Chen. 1999. B7-H1, a third member of the H. Jeon, K. H. Chung, W. K. Song, et al. 2013. Swiprosin-1 is a novel actin B7 family, co-stimulates T-cell proliferation and interleukin-10 secretion. Nat. bundling protein that regulates cell spreading and migration. PLoS One 8: Med. 5: 1365–1369. e71626. 6. Freeman, G. J., A. J. Long, Y. Iwai, K. Bourque, T. Chernova, H. Nishimura, 29. Ritter, A. T., Y. Asano, J. C. Stinchcombe, N. M. Dieckmann, B. C. Chen, L. J. Fitz, N. Malenkovich, T. Okazaki, M. C. Byrne, et al. 2000. Engagement of C. Gawden-Bone, S. van Engelenburg, W. Legant, L. Gao, M. W. Davidson, the PD-1 immunoinhibitory receptor by a novel B7 family member leads to et al. 2015. Actin depletion initiates events leading to granule secretion at the negative regulation of lymphocyte activation. J. Exp. Med. 192: 1027–1034. immunological synapse. Immunity 42: 864–876. 7. Latchman, Y., C. R. Wood, T. Chernova, D. Chaudhary, M. Borde, I. Chernova, 30. Riedl, J., A. H. Crevenna, K. Kessenbrock, J. H. Yu, D. Neukirchen, M. Bista, Y. Iwai, A. J. Long, J. A. Brown, R. Nunes, et al. 2001. PD-L2 is a second ligand F. Bradke, D. Jenne, T. A. Holak, Z. Werb, et al. 2008. Lifeact: a versatile marker for PD-1 and inhibits T cell activation. Nat. Immunol. 2: 261–268. to visualize F-actin. Nat. Methods 5: 605–607. 8. Tseng, S. Y., M. Otsuji, K. Gorski, X. Huang, J. E. Slansky, S. I. Pai, A. Shalabi, 31. Rosendahl, A., K. Kristensson, K. Riesbeck, and M. Dohlsten. 2000. T-cell T. Shin, D. M. Pardoll, and H. Tsuchiya. 2001. B7-DC, a new dendritic cell cytotoxicity assays for studying the functional interaction between the super- by guest on September 26, 2021 molecule with potent costimulatory properties for T cells. J. Exp. Med. 193: antigen staphylococcal enterotoxin A and T-cell receptors. Methods Mol. Biol. 839–846. 145: 241–257. 9. Juneja, V. R., K. A. McGuire, R. T. Manguso, M. W. LaFleur, N. Collins, 32. Vuadens, F., N. Rufer, A. Kress, P. Corthe´sy, P. Schneider, and J. D. Tissot. 2004. W. N. Haining, G. J. Freeman, and A. H. Sharpe. 2017. PD-L1 on tumor cells is Identification of swiprosin 1 in human lymphocytes. Proteomics 4: 2216–2220. sufficient for immune evasion in immunogenic tumors and inhibits CD8 T cell 33. Vega, I. E., E. E. Traverso, Y. Ferrer-Acosta, E. Matos, M. Colon, J. Gonzalez, cytotoxicity. J. Exp. Med. 214: 895–904. D. Dickson, M. Hutton, J. Lewis, and S. H. Yen. 2008. A novel calcium- 10. Azoulay-Alfaguter, I., M. Strazza, A. Pedoeem, and A. Mor. 2015. The coreceptor binding protein is associated with tau proteins in tauopathy. J. Neurochem. programmed death 1 inhibits T-cell adhesion by regulating Rap1. J. Allergy Clin. 106: 96–106. Immunol. 135: 564–567. 34. Hagen, S., S. Brachs, C. Kroczek, B. G. Fu¨rnrohr, C. Lang, and D. Mielenz. 11. Chemnitz, J. M., R. V. Parry, K. E. Nichols, C. H. June, and J. L. Riley. 2004. 2012. The B cell receptor-induced calcium flux involves a calcium mediated SHP-1 and SHP-2 associate with immunoreceptor tyrosine-based switch motif of positive feedback loop. Cell Calcium 51: 411–417. programmed death 1 upon primary human T cell stimulation, but only receptor 35. Vazquez-Rosa, E., E. N. Rodriguez-Cruz, S. Serrano, L. Rodriguez-Laureano, ligation prevents T cell activation. J. Immunol. 173: 945–954. and I. E. Vega. 2014. Cdk5 phosphorylation of EFhd2 at S74 affects its calcium 12. Hui, E., J. Cheung, J. Zhu, X. Su, M. J. Taylor, H. A. Wallweber, D. K. Sasmal, binding activity. Protein Sci. 23: 1197–1207. J. Huang, J. M. Kim, I. Mellman, and R. D. Vale. 2017. T cell costimulatory 36. Du¨tting, S., S. Brachs, and D. Mielenz. 2011. Fraternal twins: swiprosin-1/ receptor CD28 is a primary target for PD-1-mediated inhibition. Science 355: EFhd2 and Swiprosin-2/EFhd1, two homologous EF-hand containing calcium 1428–1433. binding adaptor proteins with distinct functions. Cell Commun. Signal. 9: 2. 13. Patsoukis, N., J. Brown, V. Petkova, F. Liu, L. Li, and V. A. Boussiotis. 2012. 37. Ramesh, T. P., Y. D. Kim, M. S. Kwon, C. D. Jun, and S. W. Kim. 2009. Selective effects of PD-1 on Akt and Ras pathways regulate molecular com- Swiprosin-1 regulates cytokine expression of human mast cell line HMC-1 ponents of the cell cycle and inhibit T cell proliferation. Sci. Signal. 5: ra46. through actin remodeling. Immune Netw. 9: 274–284. 14. Peled, M., A. S. Tocheva, S. Sandigursky, S. Nayak, E. A. Philips, K. E. Nichols, 38. Kroczek, C., C. Lang, S. Brachs, M. Grohmann, S. Du¨tting, A. Schweizer, M. Strazza, I. Azoulay-Alfaguter, M. Askenazi, B. G. Neel, et al. 2018. Affinity L. Nitschke, S. M. Feller, H. M. Ja¨ck, and D. Mielenz. 2010. Swiprosin-1/EFhd2 purification mass spectrometry analysis of PD-1 uncovers SAP as a new controls B cell receptor signaling through the assembly of the B cell receptor, checkpoint inhibitor. Proc. Natl. Acad. Sci. USA 115: E468–E477. Syk, and phospholipase C gamma2 in membrane rafts. J. Immunol. 184: 3665– 15. Sheppard, K. A., L. J. Fitz, J. M. Lee, C. Benander, J. A. George, J. Wooters, 3676. Y. Qiu, J. M. Jussif, L. L. Carter, C. R. Wood, and D. Chaudhary. 2004. PD-1 39. Avramidou, A., C. Kroczek, C. Lang, W. Schuh, H. M. Ja¨ck, and D. Mielenz. inhibits T-cell receptor induced phosphorylation of the ZAP70/CD3zeta 2007. The novel adaptor protein Swiprosin-1 enhances BCR signals and con- signalosome and downstream signaling to PKCtheta. FEBS Lett. 574: 37–41. tributes to BCR-induced apoptosis. Cell Death Differ. 14: 1936–1947. 16. Nishimura, H., M. Nose, H. Hiai, N. Minato, and T. Honjo. 1999. Development 40. Thylur, R. P., Y. D. Kim, M. S. Kwon, H. M. Oh, H. K. Kwon, S. H. Kim, of lupus-like autoimmune diseases by disruption of the PD-1 gene encoding an S. H. Im, J. S. Chun, Z. Y. Park, and C. D. Jun. 2009. Swiprosin-1 is expressed in ITIM motif-carrying immunoreceptor. Immunity 11: 141–151. mast cells and up-regulated through the protein kinase C beta I/eta pathway. 17. Nishimura, H., T. Okazaki, Y. Tanaka, K. Nakatani, M. Hara, A. Matsumori, J. Cell. Biochem. 108: 705–715. S. Sasayama, A. Mizoguchi, H. Hiai, N. Minato, and T. Honjo. 2001. Autoim- 41. Zhang, S., Y. Tu, Y. M. Sun, Y. Li, R. M. Wang, Y. Cao, L. Li, L. C. Zhang, and mune dilated cardiomyopathy in PD-1 receptor-deficient mice. Science 291: Z. B. Wang. 2018. Swiprosin-1 deficiency impairs macrophage immune response 319–322. of septic mice. JCI Insight DOI: 10.1172/jci.insight.95396.